No doubt the earliest observers of the needle believed that it pointed directly to the North. If such were indeed the fact the entire science of navigation would be vastly simpler than it is. But it required no very acute powers of observation to discover that the magnetized needle does not in reality point directly towards the earth's poles. There are indeed places on the earth where it does so point, but in general it is observed to deviate by a few degrees from the exact line of the meridian. Such deviation is technically known as magnetic declination. That this declination is not the same for all places was discovered by Columbus in the course of his first transatlantic voyage.
A century or so later, the accumulated records made it clear that declination is not a fixed quantity even at any given place. An Englishman, Stephen Burrows, is credited with making the discovery that the needle thus shifts its direction slightly with the lapse of time, and the matter was more clearly determined a little later by Gillebrand, Professor of Geometry at Graham College. Dr. Halley, the celebrated astronomer--whose achievements have been recalled to succeeding generations by the periodical return of the comet that bears his name--gave the matter attention, and in a paper before the Royal Society in 1692 he pointed out that the direction of the needle at London had changed in a little over a century (between 1580 and 1692) from 11 degrees 15 minutes East to 6 degrees West, or more than 17 degrees.
Halley conclusively showed that similar variations occurred at all other places where records had been kept. He had already demonstrated, a few years earlier, that the deviations of the compass noted at sea are not due to the varying attractions of neighboring bodies of land, but to some influence having to do with the problem of terrestrial magnetism in its larger aspects. Halley advocated the doctrine, which had first been put forward by William Gilbert, that the earth itself is a gigantic magnet, and that the action of the compass is dependent upon this terrestrial source and not, as many navigators believed, on the influence of a magnetic star, or on localized deposits of lodestone somewhere in the unknown regions of the North.